Chemical Engineering Communications

ISSN: 0098-6445 (Print) 1563-5201 (Online) Journal homepage: www.tandfonline.com/journals/gcec20

Thermodynamic Equilibrium Analysis of Propane

Dehydrogenation with Carbon Dioxide and Side
Reactions

Farnaz Tahriri Zangeneh, Abbas Taeb, Khodayar Gholivand & Saeed

Sahebdelfar

To cite this article: Farnaz Tahriri Zangeneh, Abbas Taeb, Khodayar Gholivand & Saeed
Sahebdelfar (2016) Thermodynamic Equilibrium Analysis of Propane Dehydrogenation with
Carbon Dioxide and Side Reactions, Chemical Engineering Communications, 203:4, 557-565,
DOI: 10.1080/00986445.2015.1017638

To link to this article: https://doi.org/10.1080/00986445.2015.1017638

Published online: 25 Jan 2016.

Submit your article to this journal

Article views: 1663

View related articles

View Crossmark data

Citing articles: 10 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gcec20
Chemical Engineering Communications, 203:557–565, 2016
Copyright # Taylor & Francis Group, LLC
ISSN: 0098-6445 print/1563-5201 online
DOI: 10.1080/00986445.2015.1017638

Thermodynamic Equilibrium Analysis of Propane

Dehydrogenation with Carbon Dioxide and Side Reactions
FARNAZ TAHRIRI ZANGENEH1,2, ABBAS TAEB1, KHODAYAR GHOLIVAND1,3, and SAEED SAHEBDELFAR2
1
Department of Chemistry, Science and Research Branch, Islamic Azad University, Tehran, Iran
2
Catalyst Research Group, Petrochemical Research and Technology Company, National Petrochemical Company, Tehran, Iran
3
Department of Chemistry, Faculty of Sciences, Tarbiat Modares University, Tehran, Iran

A thermodynamic analysis of propane dehydrogenation with carbon dioxide was performed using constrained Gibbs free energy
minimization method. Different reaction networks corresponding to different catalytic systems, including non-redox and redox
oxide catalysts, were simulated. The influences of CO2=C3H8 molar ratio (1–10), temperature (700–1000 K), and pressure (0.5–
5 bar) on equilibrium conversion and product composition were studied. In the presence of CO2 with a molar ratio of CO2=
C3H8 ¼ 1, the temperature of dehydrogenation can be 30 K lower than that of dehydrogenation in the presence of steam (H2O=
C3H8 ¼ 1) and about 50 K lower than that of simple dehydrogenation without dilution to achieve 60% propane conversion. It
was found that the occurrence of dry reforming of propane and coke-forming side reactions could strongly impact the equilibrium
product composition of the multireaction system and, therefore, these reactions should be kinetically controlled. Comparison of the
simulated reactant conversions with those reported in the literatures revealed that the experimental conversion levels of propane are
far below the corresponding equilibrium values due to rapid catalyst deactivation by coke, implying that research efforts should be
directed toward formulation of more active and selective catalysts.
Keywords: Carbon dioxide; Chemical equilibrium; Greenhouse gases; Oxidative dehydrogenation; Propane reforming; Propylene

Introduction as side reaction; therefore, proper reaction conditions are

essential for selective dehydrogenation to occur.
Propylene is an important raw material in the production of To reduce catalyst deactivation, coke-removing agents
polypropylene, polyacrylonitrile, acrolein, and acrylic acid. such as steam and hydrogen have been used as co-feeds in
The conventional manufacturing methods such as catalytic certain commercial processes. Steam controls the reaction
and steam cracking which produce propylene as a temperature, serving as heat sink and enhances the
by-product do not meet the ever-increasing demand. selectivity.
Therefore, propane dehydrogenation (PDH) has received By removing the hydrogen in the form of water from
much attention (Heinritz–Adrian et al., 2008): reaction 1, oxidative dehydrogenation proceeds to a greater
0
extent than simple catalytic dehydrogenation. Catalytic oxi-
C3 H8 , C3 H6 þ H2 DH298 ¼ 124 kJ=mol ð1Þ dative dehydrogenation of propane by oxygen is an attract-
ive alternative because it is an exothermic and
Direct catalytic dehydrogenation of hydrocarbons is charac-
non-thermodynamically limited reaction. However, the pro-
terized by low conversion per pass and low yields due to
pylene selectivity of the reaction remains an unsolved prob-
equilibrium limitation. The process is endothermic and the
lem owing to the over-oxidation of propane (Xu et al., 2006).
thermodynamic equilibrium is unfavorable (Olah and
On the other hand, soft oxidants such as carbon dioxide
Molnar, 2003). Temperatures above 800 K are required for
appear as potential alternatives. The PDH in the presence
commercially feasible conversions. At these temperatures,
of carbon dioxide (CO2–PDH) is also of interest because it
undesirable side reactions limit the propylene yield and coke
utilizes the greenhouse gas CO2 (Zangeneh et al., 2011).
formation necessitates frequent catalyst regeneration. The
CO2 is a coke-removing agent as well. Bartholomew
use of too high contact time or temperature causes cracking
(1984) reported that oxidizability in the gasification of
coke was in the following order: O2(105) > H2O(3) >
CO2(1) > H2(0.003), where numbers in parentheses indicate
the relative ratios of the gasification with oxidants. CO2
Address correspondence to Farnaz Tahriri Zangeneh, Catalyst has been called as nontraditional, unusual mild oxidant,
Research Group, Petrochemical Research and Technology Com- and oxygen transfer agent. In this sense, carbon dioxide
pany, National Petrochemical Company, Tehran, Iran. E-mail: could be proposed as an oxidant in the oxidative conversions
[email protected] of hydrocarbons.
558 F. T. Zangeneh et al.

The promoting effect of CO2 is explained either by con- PDH with carbon dioxide is characterized by rapid cata-
sumption of the hydrogen in the reverse water–gas shift reac- lyst deactivation due to coke formation. Ga2O3 showed
tion (RWGS) (reaction 2), which shifts the equilibrium of drastic deactivation during a few hours on stream (Zheng
PDH to completion (two-step pathway) or its participation et al., 2005). Besides coke formation, the reduction of Cr(VI)
in direct oxidation of propane to propylene (reaction 3) to Cr(III) is responsible for deactivation of chromium-based
(one-step pathway): catalysts (Zhang et al., 2011).
Any combination of reactions 1–3 and 6–9 could be
0 favored according to the catalyst formulation and operating
H2 þ CO2 , CO þ H2 O DH298 ¼ 40 kJ=mol ð2Þ
conditions and, accordingly, different sets of equilibria could
C3 H8 þ CO2 , C3 H6 þ CO þ H2 O 0
DH298 ¼ 164 kJ=mol be approached. A suitable catalyst should favor only the
desirable reaction(s).
ð3Þ Thermodynamic analysis of possible reactions for a series
The oxide catalysts which are active in the non-oxidative of feed components is an important step in the design of suit-
dehydrogenation of propane are also active in the process able catalyst for a particular reaction (Richardson, 1989). It
with CO2. Many bulk and supported oxide-based materials provides a basis for experimental and computational studies,
containing active species, such as Cr (Michorczyk and as well as understanding the boundaries and constraints that
Ogonowski, 2003; Michorczyk et al., 2008, 2010, 2011b), thermodynamics imposes on process and catalyst develop-
Ga (Michorczyk and Ogonowski, 2003; Xu et al., 2006), V ment (Mawson et al., 1993). The thermodynamics of CO2
(Takahara et al., 2005; Ogonowski and Skrzynska, 2008), PDH has been studied earlier (Ogonowski and Skrzynska
and Fe (Michorczyk et al., 2004) have been studied as cata- 2008; Michorczyk et al., 2011a). However, a detailed study
lysts for dehydrogenation of propane in the presence of CO2. incorporating the effect of different co-feeds and the effect
Among the investigated materials, those containing Cr or of individual side reactions is scarce.
Ga were found to be the most active and selective In this work, results are presented for a detailed computa-
(Michorczyk et al., 2011b). tional study of the chemical equilibria that apply to the
The one-step mechanism has been proposed for redox CO2–PDH. The chemical equilibria of one- and two-step
oxides (Cr2O3, Cr2O5, ZnO, and V2O5), where CO2 can par- oxidative PDH with CO2, in the absence and presence of
ticipate as an oxidizing agent in the following redox cycle competing side reactions, are studied. The equilibrium com-
(Michorczyk et al., 2008): positions were calculated by Gibbs free energy minimization
using Lagrange undermined multiplier method. The result-
C3 H8 þ MeOx , C3 H6 þ MeOx1 þ H2 O ð4Þ ing set of equations was solved using Matlab software.
The results were compared with the available experimental
CO2 þ MeOx1 , CO þ MeOx ð5Þ data in the literature.

Redox oxides involve the transfer of electron between spe- Simulation and Modeling Method
cies by different oxidation states. The two-step pathway
has been proposed for non-redox oxides such as Ga2O3 Equilibrium compositions and conversions of a reacting sys-
(Xu et al., 2006; Zheng et al., 2005). Production of H2 could tem can be obtained by minimization of the total Gibbs
be an indication of the two-step pathway. energy Gt according to Equation (10) (Perry et al., 1999).
Unfortunately, depending on the system conditions,
many undesired side reactions can occur including dry t
dGT;P ¼0 ð10Þ
reforming of propane to synthesis gas
For gas–phase reactions, the problem is to find the equilib-
0 rium composition for given T and P and for a given initial
C3 H8 þ 3CO2 , 6CO þ 4H2 DH298 ¼ 620 kJ=mol ð6Þ
feed. The objective function is subjected to the constraining
and propane decomposition, which bring about coke material balance equations, based on conservation of the
formation total number of atoms of each element in a system com-
prised of w elements. The material balance for the element
0
C3 H8 , 3C þ 4H2 DH298 ¼ 106 kJ=mol ð7Þ k is then:

Coke can also be formed by carbon dioxide disproportiona- X

ni aik ¼ Ak ðk ¼ 1; 2; :::wÞ ð11Þ
tion (Boudouard reaction, 8) and carbon monoxide i
reduction by hydrogen (reaction 9):
or
0
2CO , C þ CO2 DH298 ¼ 171 kJ=mol ð8Þ X
ni aik  Ak ¼ 0 ðk ¼ 1; 2; :::wÞ ð12Þ
0
CO þ H2 , C þ H2 O DH298 ¼ 131 kJ=mol ð9Þ i

The reverse of reactions 8 and 9 serve as coke-removing where Ak is the total number of atomic masses of the kth
reactions under appropriate conditions. element in the feed and aik is the number of atoms of the
Thermodynamic Equilibrium Analysis of Propane Dehydrogenation 559

kth element present in each molecule of chemical species i. standard Gibbs function of formation of solid carbon, and
Multiplying each constraint by kk, a Lagrange multiplier, mole of carbon, respectively.
and summing over k, the Lagrange function F is obtained The above formulation considers all possible reactions
by addition of this sum to Gt: among the reactive species (i.e., reactants or products). This
! is appropriate for noncatalytic reactions at high tempera-
X X tures where all reactions can proceed. However, in catalytic
t
F ¼G þ kk ni aik  Ak ð13Þ systems, the catalyst may promote certain reaction(s) among
k i possible reactions and change the equilibrium composition
within the time frame of interest. This can be implemented
The minimum value of both F and Gt is found when the
by specifying the reacting species and=or additional con-
partial derivatives of F with respect to ni are set equal to
straints (to the elemental balances) in the above formula-
zero. Following the necessary algebraic manipulations, it
tions (Table I). Each additional constraint (if any) will
can be shown that for gaseous systems (Michorczyk et al.,
introduce a new term and k into Equation (13), with neces-
2011a):
sary modifications in subsequent equations.
The required thermodynamic functions were obtained
yi /^i P X from the literature (Yaws, 1999), and the resulting set of
DGfi0 þ RT ln þ kk aik ¼ 0 ði ¼ 1; 2; :::NÞ ð14Þ
P0 nonlinear equations was solved by Matlab software using
k
a quasi-Newton method.
where P0 is the standard pressure, /^i and yi are the fugacity
coefficient and mole fraction of species i, respectively. For Results and Discussion
low pressures or high temperatures, as is the case for the
present work, ideal gas assumption is adequate. According to the set of reactions considered for equilibrium
There are N equilibrium equations (Equation (14)), one calculations, very different equilibrium conversions and
for each chemical species, and there are w material balance compositions could be obtained for a given feed and operat-
equations (Equation (12)), one for each element (a total of ing conditions. In this section, the results of equilibrium cal-
N þ w equations). The choice of a set of species is entirely culations for different sets of reactions are presented and
equivalent to the choice of a set of independent reactions discussed. First of all, the one- and two-step CO2–PDH
among the species. In any event, a set of species or an results without side reactions are presented and compared.
equivalent set of independent reactions must always be In subsequent sections, propane dry reforming and coke for-
assumed and different assumptions produce different results mation are further considered in the reaction network.
(Perry et al., 1999).
When solid carbon (graphite) is present, the solid–vapor
One- and Two-Step CO2 Propane Dehydrogenation
equilibrium for free energy of carbon is used as Equation
(15). Substituting Equation (13) by (14) for gaseous species For one-step reaction system, the number of chemical spe-
and by (15) for solid species, the function for free energy cies (assuming no hydrogen fed) in the gaseous phase is 5
minimization is as Equation (16) (Nikoo and Amin, 2011) with one equilibrium reaction (reaction 3). Therefore, the
number of independent components is 4, and according to
 CðgÞ ¼ G
 CðSÞ ¼ GCðSÞ ffi DG 0 the phase rule, the degree of freedom is 4. For two-step reac-
G fCðSÞ ¼ 0 ð15Þ
tion, the number of compounds is 6 with two equilibrium
! reactions, giving the same variance of 4. Consequently, by
X
N
yi /^ P X specifying T and P, the molar fraction of only two com-
ni DGfi0 þ RT ln 0i þ kk aik 0
þ ðnc DGfCðSÞ Þ pounds can vary independent of others, that is, the others
i¼1
P k may be calculated by the thermodynamic constants of the
¼0 ð16Þ reactions for both systems.
Figure 1 shows the effect of temperature on equilibrium
in which G  CðSÞ , GC(S), DG 0 , and nc are the partial
 CðgÞ , G propane conversion with different diluting agents. The con-
fCðSÞ
molar Gibbs free energy of gaseous carbon, that of solid car- versions increase with temperature because of the endo-
bon, the molar Gibbs free energy of solid carbon, the thermic nature of the overall reactions. The simple

Table I. Components and additional constraints for different reaction systems

Reaction system Components Additional constraint

Simple (reaction 1) C3H8, C3H6, H2 (or inert diluent) None

One-step (reaction 3) C3H8, CO2, C3H6, CO, H2O ðnC3 þ nC3 ¼ Þinitial feed ¼ ðnC3 þ nC3 ¼ Þequilibrium
Two-step (reaction 1 þ reaction 2) C3H8, CO2, C3H6, CO, H2, H2O ðnC3 þ nC3 ¼ Þinitial feed ¼ ðnC3 þ nC3 ¼ Þequilibrium
Two-step þ propane reforming C3H8, CO2, C3H6, CO, H2, H2O None
Two-step þ propane reforming þ coke formation C3H8, CO2, C3H6, CO, H2, H2O, C None
560 F. T. Zangeneh et al.

Figure 2 shows the effect of diluent concentration at 873

and 923 K. Except for hydrogen diluted system, an increase
in diluent concentration will increase the equilibrium pro-
pane conversion, with the highest conversion being observed
for the two-step CO2–PDH. The one-step CO2–PDH, on the
contrary, shows the lowest conversion at lower diluent con-
centrations and even no conversion at zero CO2 concen-
tration. However, the non-redox oxides catalyzing CO2–
PDH reaction catalyze simple dehydrogenation in the
absence of CO2 as well, and consequently, propane conver-
sion should be observed even at zero CO2 concentration.
Therefore, one-step CO2–PDH equilibrium, by itself, could
not be a good representative of the real systems especially
at low CO2 concentrations, and its combination with simple
dehydrogenation should be a better approximation of the
real systems. These results are also in good agreement with
Fig. 1. The effect of temperature on equilibrium conversion of those obtained by Michorczyk and Ogonowski (2003).
propane for diluent=propane molar ratio of unity at 1 bar (no One might note that the combination of one-step and sim-
dilution in the case of simple dehydrogenation, dashed line) ple dehydrogenation gives exactly the same results as
((&) steam diluted, (&) hydrogen diluted, (~) one-step CO2– two-step CO2–PDH since reactions 1–3 are not thermodyna-
PDH, and (4) double-step CO2–PDH and one-stepþsimple mically independent. This is not the case, however, when the
PDH). reactions are kinetically controlled, i.e., when the reactions
proceed independently by their own mechanism.
Figure 3 shows the effect of total pressure on equilibrium
dehydrogenation (i.e.,, without diluent) is also depicted as propane conversion at 873 and 923 K. The higher the press-
the reference. Addition of a diluent (both inert and=or ure, the lower is the propane conversion in all cases as the
reactive) increases the conversion through decreasing the overall conversion is accompanied with a net increase of
partial pressure of the products, and also, in the case of car- the total number of moles.
bon dioxide, consuming the produced hydrogen. In the case Table II shows the temperatures required for 50% and
of hydrogen co-feed, however, the effect is more complex. 60% propane conversions (typical commercial values) for
Hydrogen addition tends to increase the conversion through different pathways. For 60% conversion, equimolar addition
decreasing partial pressure, as other diluents, and to of CO2 to the feed reduces the reaction temperature by 54
decrease the equilibrium conversion by shifting reaction 1 and 29 K compared to non-diluted and steam diluted sys-
to the left as H2 is a product. The net effect is a slight drop tems, respectively, illustrating the advantage of using CO2
in conversion as illustrated in Figure 1. The highest conver- as a mild oxidant in PDH.
sion is observed in the case of two-step CO2–PDH. These
results are closely related to the results obtained for dehy-
drogenation of propane (Michorczyk and Ogonowski, Competitive Reforming of Propane (and/or Propylene)
2003) and ethylbenzene to styrene (Qin et al., 2003; Sun In the previous section, only combinations of reactions
et al., 2004). involved in direct and indirect CO2–PDH (i.e., reactions

Fig. 2. The effect of diluent concentration on equilibrium propane conversion at P ¼ 1 bar and (a) 873 and (b) 923 K ((&) steam
diluted, (&) hydrogen diluted, (~) one-step CO2–PDH, and (4) double-step CO2–PDH and one-stepþsimple PDH).
Thermodynamic Equilibrium Analysis of Propane Dehydrogenation 561

Fig. 3. The effect of total pressure on equilibrium propane conversion for diluent=C3H8 ¼ 1 in initial feed at (a) 873 and (b) 923 K
(no dilution in the case of simple dehydrogenation, dashed line) ((&) steam diluted, (&) hydrogen diluted, (~) one-step CO2–PDH,
and (4) double-step CO2–PDH, and one-stepþsimple PDH).

1–3) were considered for chemical equilibrium calculations. temperature. The subsequent decrease in equilibrium carbon
The results are only applicable to systems with highly selec- monoxide concentration is simply due to the increase of total
tive catalysts. An important side reaction could be dry moles resulted from the formation of lower-molecular-
reforming of propane (reaction 6). The degree of freedom weight compounds at higher temperatures.
of the system is 4 from the phase rule. Two compositions Figure 6 shows the effect of carbon dioxide concentration
in addition to T and P are required to specify the system. (yCO2) in the initial feed on the equilibrium conversion of
Figure 4 shows the effect of the temperature on the equi- propane and carbon dioxide and selectivity to propylene at
librium conversion of the reactants and propylene selectivity 873 K when propane reforming is the competing reaction.
for equimolar C3H8=CO2 feed at atmospheric pressure when At very low CO2 contents in the feed, the product distri-
propane reforming is the competing reaction. The conver- bution is the same as that of simple PDH. The conversion
sion of CO2 is much higher than that of propane, implying of propane increases with CO2 mole fraction at the expense
a rather low extent of PDH reaction. Both propane and car- of propylene selectivity. The equilibrium CO concentration
bon dioxide conversions increase with temperature due passes a maximum at CO2 mole fraction of about 0.75
to the exothermic nature of the main reactions. At higher (Figure 7) corresponding to the stoichiometric value for pro-
temperatures, the conversion of both reactants is nearly pane dry reforming. This is reasonable as dry reforming of
complete, while the equilibrium propylene selectivity propane is the main carbon dioxide consuming reaction at
approaches 0.67.
Figure 5 shows the effect of temperature on equilibrium
product compositions. The equilibrium compositions of
both CO and H2 approach asymptotically to 0.429, while
those of propylene and water to 0.153 and zero, respectively.
This illustrates that at higher temperatures, CO2 is involved
in thermodynamically more favorable dry reforming of pro-
pane, while propylene is obtained via simple dehydrogena-
tion of the excess propane.
Around 823 K, carbon dioxide conversion is nearly com-
plete and the mole fraction of CO (the main oxygen-
containing component present) passes a maximum at this

Table II. Temperatures required for 50% and 60% conversion

(diluent=C3H8 ¼ 1, P ¼ 1 bar)
Temperature, K

Conversion No CO2 CO2

(%) dilution Steam Hydrogen (one-step) (two-step) Fig. 4. The effect of temperature on equilibrium conversion of
propane and carbon dioxide and selectivity of propylene in
50 875 850 905 891 821 the presence of propane reforming reaction (C3H8=CO2 ¼ 1,
60 901 876 928 930 847 P ¼ 1 bar).
562 F. T. Zangeneh et al.

Fig. 5. The effect of temperature on equilibrium product

composition in the presence of propane reforming reaction Fig. 7. The effect of carbon dioxide mole fraction on equilib-
(C3H8=CO2 ¼ 1, P ¼ 1 bar). rium product composition in the presence of propane reforming
reaction (T ¼ 873 K, P ¼ 1 bar).

this temperature. However, when CO2 concentration passes

the stoichiometric ratio and no propane is available, part of are widely used and commercialized for PDH have not been
the excess CO2 can react through reverse water gas shift successfully used and not reported for carbon dioxide PDH.
reaction to form water. The water concentration passes a These catalysts, like most other Group VIII metals, are good
maximum around yCO2 of 0.9 where both CO2 and H2 con- catalysts for dry reforming of propane (Solymosi et al.,
centrations are sufficiently high to form water. 2005). Consequently, the selectivity to propylene should be
It should be pointed out that as steam reforming of pro- unacceptably low over Pt-based catalysts.
pane (another possible side reaction) is not thermodynami-
cally independent of the set of reactions used in Coke Formation
calculations; its effect on equilibrium conversions and com-
positions has been inherently considered. Coke formation is characteristic of many catalytic reactions
In summary, the results indicate that under typical CO2– involving hydrocarbons and other carbon-containing feeds
PDH conditions, propane dry reforming is thermodynami- such as carbon dioxide. It reduces the catalyst activity and
cally much more favored, making the product distribution lifetime. Therefore, its inclusion in thermodynamic equilib-
unfavorable. This can explain why Pt-based catalysts which rium calculations provides a good assessment on the suitable
reaction conditions for stable operation.

Fig. 6. The effect of carbon dioxide mole fraction on equilib- Fig. 8. The effect of temperature on equilibrium conversion of
rium conversion of propane and carbon dioxide and selectivity propane and carbon dioxide and selectivity of propylene in
of propylene in the presence of propane reforming reaction the presence of propane reforming and coke-forming reactions
(T ¼ 873 K, P ¼ 1 bar). (C3H8=CO2 ¼ 1, P ¼ 1 bar).
Thermodynamic Equilibrium Analysis of Propane Dehydrogenation 563

Fig. 9. The effect of temperature on equilibrium product com- Fig. 10. The effect of carbon dioxide mole fraction on equilib-
position in the presence of propane reforming and coke-forming rium conversion of propane and carbon dioxide and selectivity
reactions (C3H8=CO2 ¼ 1, P ¼ 1 bar). of propylene in the presence of propane reforming and
coke-forming reactions (T ¼ 873 K, P ¼ 1 bar).

When coke-forming reactions are included in equilibrium

propane content in the initial feed mixture approaches zero
calculations, the product compositions change dramatically.
the overall reaction stoichiometry is approximated by the
The degree of freedom is 3 for the system. Figure 8 shows the
following reaction
effect of temperature on the equilibrium conversions and
propylene selectivity for equimolar C3H8=CO2 feed at
atmospheric pressure when both propane reforming and C3 H8 þ 3CO2 , 2C þ 2H2 þ 2H2 O þ 4CO ð17Þ
coke formation are the competing reactions. The conversion
of propane is nearly complete and the selectivity to propy- which is a combination of reactions 6 and 9.
lene is negligibly small. The conversion of carbon dioxide These results show that coke formation is also thermody-
is relatively high over the range shown. namically much more favorable than CO2–PDH and high-
The main carbon-containing product is coke, the extent light again the critical role of catalyst for successful
of which decreases with increasing temperature (Figure 9). implementation of CO2–PDH by kinetic control of side reac-
This can be attributed to the fact that coke gasification reac- tions. As a consequence, acidic supports or oxides are not
tions (the reverse of reactions 8 and 9) converting coke to desirable as they favor cracking and coke formation (i.e.,
CO are endothermic and are favored at higher temperatures. acid coking (Richardson, 1989)). In fact, basic oxides have
At higher temperatures (>1000 K), the mole fraction of both
coke and CO in products approaches to 0.25 and that of
hydrogen to 0.5, with other components being present in
very small amounts. The main reactions are thus coke for-
mation and propane dry reforming.
Figure 10 shows the effect of carbon dioxide concen-
tration in the initial feed on the equilibrium conversion of
propane and carbon dioxide and selectivity to propylene at
873 K when propane reforming and coke formation are the
competing reactions. The conversion of propane is nearly
complete over the whole range shown while that of carbon
dioxide decreases monotonically with increasing CO2 con-
tent in the feed. The selectivity to propylene is negligible.
The main carbon-containing product is coke over most of
the range (Figure 11). At very low CO2 mole fractions, the
main reaction is propane pyrolysis. Upon increasing yCO2,
the selectivity to both H2 and coke (as well as their equilib-
rium mole fraction) decreases which is due to their oxidation
by CO2. Water and CO mole fractions pass peaks at yCO2 of
ca. 0.7 and 0.9, respectively. At very high CO2 concentra- Fig. 11. The effect of carbon dioxide mole fraction on equilib-
tions, the mole fractions of H2O, H2, and coke become ident- rium product composition in the presence of propane reforming
ical but smaller than that of CO, indicating that when and coke-forming reactions (T ¼ 873 K, P ¼ 1 bar).
564 F. T. Zangeneh et al.

Table III. A comparison of experimental (X) and equilibrium conversions (Xe) for simple PDH under the same conditions

Feed (molar ratio) Temperature Pressure

Catalyst C3=H2=N2=Steam (K) (bar) TOS (h) X S Xe References

Pt–Sn=hydrotalcite 32=4.5=24=38.5 873 1 5 0.59 0.93 0.61 Akporiaye et al. (2001)

Pt=h–Al2O3 1=0=0=4 823 1 NA 0.50 0.90 0.53 Kogan et al. (2001)
Pt=h–Al2O3 1=4=0=0 823 1 NA 0.089 0.95 0.12 Kogan et al. (2001)
Pt–Sn=MgAl2O4 1=0=0=2 823 1 NA 0.39 0.95 0.45 Dong et al. (1997)
Cr2O3=Al2O3 1=0=0=0 833 1 0.17 0.35 0.89 0.35 Pakhomov et al. (2009)
Cr2O3=Al2O3 1=0=0=0 853 1 0.17 0.39 0.80 0.42 Pakhomov et al. (2009)
Cr2O3=Al2O3 1=0=0=0 863 1 0.17 0.47 0.98 0.46 Pakhomov et al. (2009)

NA: not available, S: propylene selectivity.

Table IV. A comparison of experimental (X) and (two-step) equilibrium conversions (Xe) at p ¼ 1 bar

Catalyst C3H8=CO2=inert Temperature, K X C3 H8 XeC3H8 XCO2 XeCO2 References

Ga2O3=TiO2a 1=2=37 873 32 96.70 30 25.74 Xu et al. (2006)

Ga2O3=ZrO2a 1=2=37 873 30 96.70 29 25.74 Xu et al. (2006)
0.04CrSBA1a 1=5=9 873 39.4 95.0 6.2 13.58 Michorczysk et al. (2010)
0.04CrSBA1a 1=5=9 923 55.2 98.22 9.9 14.71 Michorczysk et al. (2010)
0.05Cr2O3=MSU–x 1=3=0 823 32.8 67.42 13.8 14.29 Liu et al. (2007)
0.05Cr2O3=MSU–x 1=3=0 873 50.4 83.89 20.7 17.97 Liu et al. (2007)
a
Reaction time: 10 min.

been favored for CO2–PDH reaction due to the acidic nature rapid catalyst deactivation by coke formation, rendering
of CO2. However, the alkali and alkaline earth oxides are inef- precise conversion measurements impossible. To reduce this
ficient because they form carbonate species in the presence of effect, often a higher than stoichiometric ratio of CO2=C3H8
carbon dioxide (Rangel et al., 2012). Metal oxides of moderate is employed in experimental works and the measurements
basicity are required and must participate in the redox process are performed in relatively short times-on-stream.
with the reduction of carbon dioxide (Krylov et al., 1995). Propane and propylene are both thermodynamically
unstable under typical dehydrogenation conditions. Conse-
Applicability to Practical Systems quently, the coke-removing capability of the diluent plays
an important role on catalyst stability. Carbon dioxide is a
The results of the previous sections imply that equilibrium weaker coke-removing agent when compared to steam.
compositions obtained in the presence of propane reforming The fact that PDH with H2 (which is a weaker
and coke-forming side reactions are of little applicability to coke-removing diluent compared to CO2) has been commer-
practical catalytic systems. Although these reactions are lar- cialized illustrates that the current catalyst formulations can-
gely controlled by the catalyst formulation, their occurrence not effectively control coke formation in CO2–PDH.
(especially coke formation), even to a small extent, can sig- Consequently, much is to be done to achieve active and
nificantly affect the catalyst performance in CO2–PDH by coke-resistant catalysts to make use of the advantages of car-
covering and modifying the active sites. bon dioxide in commercial dehydrogenation of light alkanes.
For proper comparison of thermodynamic simulation
results with experimental data, the experimental system
should be close to equilibrium-controlled regime for the Conclusions
reaction(s) under consideration. Therefore, among experi-
mental data reported for a given operating conditions over The effect of operating conditions and side reactions on the
different catalysts, only those with the highest conversion equilibrium conversions and product compositions in
(and selectivity, if applicable) were selected for comparison. oxidative dehydrogenation of propane with carbon dioxide
Table III shows a comparison of propane conversions for was studied. The inclusion of different side reactions such
simple dehydrogenation (hydrogen or inert diluted) using vari- as propane reforming and coke formation which corre-
ous diluent combinations over different catalysts with those sponds to different catalysts and operating conditions in
obtained from equilibrium calculation. A good agreement is the reaction network results in dramatic changes in equilib-
observed in most cases, indicating that excellent results (close rium compositions. Some generalization of the results could
to equilibrium) could be achieved for simple PDH in practice. be outlined as below:
Unfortunately, this is not the case for carbon dioxide . Thermodynamic considerations showed that the equilib-
dehydrogenation, as shown in Table IV. The main cause is rium conversions of propane in the presence of CO2 are
Thermodynamic Equilibrium Analysis of Propane Dehydrogenation 565

higher than in the presence of an inert gas or hydrogen Michorczyk, P., Ogonowski, J., and Niemczyk, M. (2010). Investi-
with two-step > one-step. gation of catalytic activity CrSBA-1 materials obtained by direct
. The reported experimental conversion levels for one- and method in the dehydrogenation of propane with CO2, Appl. Catal.
A: Gen., 374, 142–149.
two-step PDH are much lower than those predicted by Michorczyk, P., Niekurzak, R., Ogonowski, J., and Jasiski, R. (2011a).
equilibrium which could be attributed to the rapid catalyst Dehydrogenation of propane with CO2-experimental and
deactivation. thermodynamic considerations, Khim. Khim. Tekhnol., 54,
. The inclusion of side reactions in thermodynamic simula- 82–85.
tion greatly increases propane conversion and lowers pro- Michorczyk, P., Ogonowski, J., and Zenczak, K. (2011b). Activity of
pylene selectivity as they are thermodynamically more chromium oxide deposited on different silica supports in the dehy-
drogenation of propane with CO2 – a comparative study, J. Mol.
favorable than the main reaction.
Catal. A: Chem., 349, 1–12.
. Higher temperatures enhance the product selectivity to Nikoo, K. M., and Amin, N. A. S. (2011). Thermodynamic analysis of
CO and hydrogen, that is, propane dry reforming. carbon dioxide reforming of methane in view of solid carbon forma-
. Higher temperatures and CO2=C3H8 ratios reduce coke for- tion, Fuel Process. Technol., 92, 678–691.
mation tendency and propylene yield as well, except for cat- Ogonowski, J., and Skrzynska, E. (2008). Conversion of lower hydro-
alysts controlling undesired side reactions kinetically. carbons in the presence of carbon dioxide: The theoretic analysis
. Further research and development works are necessary to and catalytic tests over active carbon supported vanadium oxide,
Catal. Lett., 124, 52–58.
develop more efficient catalysts to realize this environmen-
Olah, G. A., and Molnar, A. (2003). Hydrocarbon Chemistry, Chapter 3,
tally favorable process. Wiley, New Jersey.
Perry, R. H., Green, D. W., and Maloney, J. O. (1999). Perry’s
Acknowledgments Chemical Engineers’ Handbook, McGraw-Hill, New York.
Pakhomov, N. A., Kashkin, V. N., Nemykina, E. I., Molchanov, V. V.,
Saeed Sahebdelfar thanks Professor Farhad Khorasheh Nadtochiy, V. I., and Noskova, A. S. (2009). Dehydrogenation of
from Department of Chemical and Petroleum Eingineering, C3–C4 paraffins on Cr2O3=Al2O3 catalysts in fluidized and fixed
Sharif University of Technology, for fruitful discussions. bed reactors, Chem. Eng. J., 154, 185–188.
Qin, Z., Liu, J., Sun, A., and Wang, J. (2003). Reaction coupling in the
new processes for producing styrene from ethylbenzene, Ind. Eng.
References Chem. Res., 42, 1329–1333.
Akporiaye, D., Jensen, S. F., Olsbye, U., Rohr, F., Rytter, E., Rangel, M. C., Monteiro, A. P. M., Oportus, M., Reyes, P., Ramos, M.
Rønnekleiv, M., and Spjelkavik, A. I. (2001). A novel, highly S., and Lima, S. B. (2012). Ethylbenzene dehydrogenation in the
efficient catalyst for propane dehydrogenation, Ind. Eng. Chem. presence of carbon dioxide over metal oxides, in G. Liu (Ed.),
Res., 40, 4741–4748. Greenhouse Gases–Capturing, Utilization and Reduction, ISBN:
Bartholomew, C. H. (1984). Catalyst deactivation, Chem. Eng., 91, 978–953–51–0192–5, InTech, doi: 10.5772=33255. Available from:
96–112. http://www.intechopen.com/books/greenhouse-gases-capturing-
Dong, W.-S., Wang, X.-K., and Peng, S.-Y. (1997). Effect of promoters on utilization-and-reduction/ethylbenzene-dehydrogenation-in-the-
catalytic performance of Pt–Sn=MgAl2O3 catalyst for propane dehy- presence-of-carbon-dioxide-over-metal-oxides-
drogenation in the presence of steam, J. Nat. Gas Chem., 6, 37–43. Richardson, J. T. (1989). Principles of Catalyst Development, Plenum
Heinritz-Adrian, M., Wenzel, S., and Youssef, F. (2008). Advanced Press, New York, 49–82.
propane dehydrogenation, Petrol. Tech. Quart., 13, 83–92. Solymosi, F., Tolmacsov, P., and Kedves, K. (2005). CO2 reforming of
Kogan, S. B., Schramm H., and Herskowitz, M. (2001). Dehydrogena- propane over supported Rh, J. Catal., 216, 377–385.
tion of propane on modified Pt=h–alumina Performance in hydro- Sun, A., Qin, Z., Chen S., and Wang, J. (2004). Role of carbon dioxide
gen and steam environment, Appl. Catal. A: Gen., 208, 185–191. in the ethylbenzene dehydrogenation coupled with reverse
Krylov, O. V., Mamedov, A. K., and Mirzabekova, S. R. (1995). The water–gas shift, J. Mol. Catal. A: Chem., 210, 189–195.
regularities in the interaction of alkanes with CO2 on oxide Takahara, I., Saito, M., Inaba, M., and Murata, K. (2005). Dehydro-
catalysts. Catal. Today, 24, 371–375. genation of propane over a silica–supported vanadium oxide cata-
Liu, L., Li, H., and Zhang, Y. (2007). Mesoporous silica–supported lyst, Catal. Lett., 102, 201–205.
chromium catalyst: Characterization and excellent performance in Xu, B., Zheng, B., Hua, W., Yue, Y., and Gao, Z. (2006). Support
dehydrogenation of propane to propylene with carbon dioxide, effect in dehydrogenation of propane in the presence of CO2 over
Catal. Commun., 8, 565–570. supported gallium oxide catalysts, J. Catal., 239, 470–477.
Mawson, S., McCutchen, S. M., Lim, P. K., and Roberts, G. W. Yaws, C. Y. (1999). Chemical Properties Handbook: Physical, Thermo-
(1993). Thermodynamics of higher alcohol synthesis, Energ. Fuel., dynamic, Environmental, Transport Safety and Health Related
7, 257–267. Properties for Organic and Inorganic Chemicals, McGraw– Hill
Michorczyk, P., and Ogonowski, J. (2003). Dehydrogenation of pro- Inc., New York.
pane in the presence of carbon dioxide over oxide-based catalysts, Zangeneh, F. T., Sahebdelfar, S., and Ravanchi, M. T. (2011). Conver-
React. Kinet. Catal. Lett., 78, 41–47. sion of carbon dioxide to valuable petrochemicals: An approach to
Michorczyk, P., Kustrowski, P., Chmielarz, L., and Ogonowski, J. clean development mechanism, J. Nat. Gas Chem., 20, 219–231.
(2004). Influence of redox properties on activity of iron oxide cata- Zhang, F., Wu, R., Yue, Y., Yang, W., Gu, S., Miao, C., Hua, W., and
lysts in the dehydrogenation of propane with CO2, React. Kinet. Gao, Z. (2011). Chromium oxide supported on ZSM–5 as a novel
Catal. Lett., 82, 121–130. efficient catalyst for dehydrogenation of propane with CO2, Micro-
Michorczyk, P., Ogonowski, J., Kustrowski, P., and Chmielarz, L. por. Mesopor. Mat., 145, 194–199.
(2008). Chromium oxide supported on MCM-41 as a highly active Zheng, B., Hua, W., Yue, Y., and Gao, Z. (2005). Dehydrogenation of
and selective catalyst for dehydrogenation of propane with CO2, propane to propylene over different polymorphs of gallium oxide, J.
Appl. Catal. A: Gen., 349, 62–69. Catal., 232, 143–151.